TextGenerationJobConfig - Amazon SageMaker
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

TextGenerationJobConfig

The collection of settings used by an AutoML job V2 for the text generation problem type.

Note

The text generation models that support fine-tuning in Autopilot are currently accessible exclusively in regions supported by Canvas. Refer to the documentation of Canvas for the full list of its supported Regions.

Contents

BaseModelName

The name of the base model to fine-tune. Autopilot supports fine-tuning a variety of large language models. For information on the list of supported models, see Text generation models supporting fine-tuning in Autopilot. If no BaseModelName is provided, the default model used is Falcon7BInstruct.

Type: String

Length Constraints: Minimum length of 1. Maximum length of 256.

Pattern: ^[a-zA-Z0-9](-*[a-zA-Z0-9])*

Required: No

CompletionCriteria

How long a fine-tuning job is allowed to run. For TextGenerationJobConfig problem types, the MaxRuntimePerTrainingJobInSeconds attribute of AutoMLJobCompletionCriteria defaults to 72h (259200s).

Type: AutoMLJobCompletionCriteria object

Required: No

ModelAccessConfig

The access configuration file to control access to the ML model. You can explicitly accept the model end-user license agreement (EULA) within the ModelAccessConfig.

Type: ModelAccessConfig object

Required: No

TextGenerationHyperParameters

The hyperparameters used to configure and optimize the learning process of the base model. You can set any combination of the following hyperparameters for all base models. For more information on each supported hyperparameter, see Optimize the learning process of your text generation models with hyperparameters.

  • "epochCount": The number of times the model goes through the entire training dataset. Its value should be a string containing an integer value within the range of "1" to "10".

  • "batchSize": The number of data samples used in each iteration of training. Its value should be a string containing an integer value within the range of "1" to "64".

  • "learningRate": The step size at which a model's parameters are updated during training. Its value should be a string containing a floating-point value within the range of "0" to "1".

  • "learningRateWarmupSteps": The number of training steps during which the learning rate gradually increases before reaching its target or maximum value. Its value should be a string containing an integer value within the range of "0" to "250".

Here is an example where all four hyperparameters are configured.

{ "epochCount":"5", "learningRate":"0.5", "batchSize": "32", "learningRateWarmupSteps": "10" }

Type: String to string map

Map Entries: Minimum number of 0 items. Maximum number of 30 items.

Key Length Constraints: Maximum length of 32.

Key Pattern: ^[a-zA-Z0-9._-]+$

Value Length Constraints: Maximum length of 16.

Value Pattern: ^[a-zA-Z0-9._-]+$

Required: No

See Also

For more information about using this API in one of the language-specific Amazon SDKs, see the following: